--On September 23, 2007 1:41:03 PM +0200 [EMAIL PROTECTED] wrote:
By the way, I have now finished an import of 6400 users into the CMS.
This is
the actual real user data. I divided it up in two steps, first imported
3000
users, then ran db_recover -c (not sure about if the -c option was
needed, but
it seemed to help before), then imported the rest 3400 and then ran
db_recover
-c again. Now Ldapbrowser lists all users in just a few seconds, and a
search
for uid=a* takes just a few seconds as well. So all seems to work now.
However
I'm wondering if the reason it's working now is that we have fewer users
(6450
now compared to ~15000 before). I will do more test imports today, in
3000-user
chunks, with db_recover -c in between, and see if/when things go haywire.
Why are you importing it in chunks?
In any case, I would still note it is highly advised to use someone else
builds than the ones RedHat provides, which you haven't indicated whether
or not you are doing now. It is likely that the RedHat BDB builds also
miss one or more required patches to BDB, which would be why you are having
problems with the transaction logs.
Just one more thing, about these transaction logs... The way I understand
it,
db_recover uses them all each time it is run. Then, when exactly is it
safe to
delete them?
I generally back up via slapcat, so I have a known point in time from which
db_recover can continue. You can set flags in the DB_CONFIG file to
automatically delete unnecessary log files.
--Quanah
--
Quanah Gibson-Mount
Principal Software Engineer
Zimbra, Inc
--------------------
Zimbra :: the leader in open source messaging and collaboration
---
You are currently subscribed to [EMAIL PROTECTED] as: [EMAIL PROTECTED]
To unsubscribe send email to [EMAIL PROTECTED] with the word UNSUBSCRIBE as the
SUBJECT of the message.
